TOEIC Link Speaking — Conversational Grounding And Clarification Strategies: The Repair-Initiation Move That Converts Comprehension Failure From Score Penalty Into Discourse-Competence Evidence
The TOEIC Link speaking rubric distinguishes the band-25-and-above performance from the band-22 ceiling on a single under-discussed dimension — the candidate's handling of the prompts in which the candidate has incompletely understood the question. The band-22 candidate, encountering a prompt that the candidate has not fully parsed, defaults to a guessed response that addresses the candidate's best reconstruction of the prompt and accepts the off-topic penalty that the rater applies when the reconstruction is wrong; the band-25 candidate executes a structured repair-initiation move that flags the comprehension gap, requests the specific clarification, and re-enters the prompt with the missing information installed. The repair-initiation move is not a deficiency signal; it is the rubric-recognized discourse-competence evidence that the rater is calibrated to reward, and the move converts the comprehension failure from a score penalty into a rubric-credit asset.
The structural difference between the two candidates is not the comprehension itself; both candidates experience the comprehension gap at comparable rates, particularly on the prompts that combine unfamiliar domain vocabulary with non-standard accent variation or with high-density discourse markers. The structural difference is the response strategy under uncertainty. The band-22 candidate treats the comprehension gap as a private problem that must not be visible to the rater and produces an externally confident response that internally is a guess; the rater detects the guess through the answer's failure to address the actual prompt and assigns the off-topic penalty regardless of how grammatically polished the guess is. The band-25 candidate treats the comprehension gap as a public discourse event that must be repaired through the rubric-recognized move and produces an externally measured response that signals the candidate's discourse competence; the rater registers the repair-initiation and assigns the discourse-competence credit regardless of whether the eventual response is grammatically simpler than the band-22 candidate's guess.
This guide formalizes the four-step conversational grounding pattern, the lexical inventory that signals repair without signaling deficiency, and the four-week installation drill that converts the move from absent to automatic. For broader speaking-strategy context, see the speaking strategic pausing and cognitive load distribution guide and the speaking self correction and repair strategies guide.
Why the silent guess caps at band 22
The TOEIC Link speaking rubric scores responses on three dimensions — task fulfillment, language use, and discourse competence. The first dimension penalizes responses that fail to address the actual prompt regardless of the response's surface quality, which is the dimension on which the silent-guess strategy fails. The third dimension rewards responses that demonstrate the candidate's command of the conversational repair mechanisms that competent speakers deploy under real-world uncertainty, which is the dimension on which the silent-guess strategy forfeits the credit that the repair-initiation strategy captures.
The rubric construction is deliberate. The test designers recognize that comprehension failure is a routine feature of real-world conversation, particularly across the accent and register variation that the TOEIC Link speaking prompts deliberately introduce. The rubric is not designed to reward the candidate who never fails to comprehend; the rubric is designed to reward the candidate who handles the comprehension failures competently when they occur. The candidate who silently guesses is signaling to the rater that the candidate has not developed the discourse-competence layer that competent speakers possess; the candidate who initiates repair is signaling that the discourse-competence layer is built and is operationally available.
The cap at band 22 is the rubric's encoded judgment that a candidate who silently guesses on the unclear prompts has demonstrated language proficiency but has not demonstrated the discourse competence that distinguishes proficient speaking from advanced speaking. The discrimination is structural — no amount of grammatical accuracy, vocabulary range, or pronunciation polish on the responses that the candidate has correctly comprehended moves the score above the band-22 ceiling if the comprehension-failure responses are handled through silent guessing. The repair-initiation move is the discourse-competence evidence that the rubric requires the candidate to produce, and the move's installation is the band-22-to-band-25 transition.
The four-step conversational grounding pattern
The repair-initiation move consists of four steps executed in a fixed sequence within the first three to four seconds of the response window. The sequence is what produces the rubric-recognizable repair structure that the rater is scanning for; departures from the sequence frequently fail to register as the rubric-rewarded move even when the propositional content of the steps is present.
Step 1 — Repair signaling
The first step is the explicit signaling that the candidate is initiating repair rather than responding. The signaling must occur within the first two seconds of the response window and must use lexical patterns that the rater is calibrated to recognize as repair initiation rather than as hesitation. Recommended signal patterns include I want to make sure I have the question right, let me check my understanding of the prompt, I want to confirm one detail before I respond, and I want to verify what you are asking. The signaling vocabulary should be drawn from the recommended patterns rather than improvised during the installation phase because the recommended patterns are the patterns that the rater has been calibrated against; using calibrated patterns ensures that the rubric credit is assigned.
The repair signaling distinguishes the rubric-rewarded move from the hesitation-filler pattern that the band-19 candidate frequently produces. The hesitation filler (um, well, let me think) signals to the rater that the candidate is buying time without a structured purpose; the repair signaling signals that the candidate is executing a structured discourse move with a defined endpoint. The structural difference is the difference between the rubric penalty that the hesitation filler accrues and the rubric credit that the repair signaling earns.
Step 2 — Comprehension articulation
The second step is the explicit articulation of the comprehension that the candidate does have, in the form of a paraphrase or restatement of the part of the prompt that the candidate has confidently parsed. The articulation should be a single complete sentence that demonstrates the candidate's partial comprehension and that localizes the gap that the candidate needs to repair. Recommended structures include I understand that you are asking about X, and I want to confirm whether Y is also part of the question, I have the part about X, but I want to make sure I have the part about Y correctly, and the question is asking about X, and the specific aspect I want to verify is Y.
The comprehension articulation is what converts the repair from an undifferentiated request for clarification into a targeted clarification request. The rater scores the targeted request as evidence that the candidate has parsed enough of the prompt to localize the gap and is asking the specific question that closes the gap; the undifferentiated request (can you repeat the question) is scored as evidence that the candidate has parsed almost none of the prompt and is asking for a global retransmission that competent speakers rarely need. The targeted-request structure is the discourse-competence signature that the rubric rewards.
Step 3 — Clarification request
The third step is the explicit request for the specific information that closes the comprehension gap, formatted as a yes-no question or as a wh-question depending on the gap type. The yes-no formatting (is the question asking about X or about Y) is appropriate when the candidate has identified the alternative interpretations and needs the prompt-author to select among them; the wh-formatting (what is the specific X that the question is asking about) is appropriate when the candidate has identified the missing component and needs the prompt-author to provide it.
The formatting choice signals the precision of the candidate's comprehension. The yes-no formatting signals that the candidate has parsed enough of the prompt to enumerate the plausible interpretations and is requesting disambiguation; the wh-formatting signals that the candidate has parsed enough of the prompt to identify the missing slot and is requesting the slot-filler. Both formats are rubric-rewarded, but the formats are not interchangeable; the candidate who uses the wh-format when the gap requires disambiguation produces a request that does not close the gap and forces a second clarification round that the rater scores as evidence of imprecise comprehension.
Step 4 — Re-entry signaling
The fourth step is the explicit signaling that the candidate is re-entering the response after the clarification, executed within one to two seconds of receiving the clarification information. Recommended signal patterns include thank you, now I understand, with that clarification, the answer is, now that I have that detail, I would respond, and given that clarification, my response is. The re-entry signaling closes the repair sequence and marks the transition from the repair-initiation discourse move to the response proper.
The re-entry signaling is the step that distinguishes the rubric-rewarded full repair sequence from the partial repair sequence that the band-22-attempting-band-25 candidate frequently produces. The partial sequence executes the first three steps and then transitions to the response without the explicit re-entry signaling, which produces a response that the rater scores as a continuation of the repair rather than as a competent post-repair answer; the full sequence executes the re-entry signaling and produces a response that the rater scores as the complete discourse-competence demonstration that the rubric rewards. The four-step discipline is what produces the rubric credit; the partial discipline is what produces the band-22 ceiling.
The lexical inventory that signals competence rather than deficiency
The repair-initiation move depends critically on the lexical choices that signal the move to the rater. The wrong lexical choices convert what should read as discourse competence into what reads as comprehension failure, and the rater scores the comprehension-failure interpretation regardless of the structural correctness of the four-step sequence. The lexical inventory is the operational layer that determines whether the repair-initiation move accrues rubric credit or rubric penalty.
The competence-signaling lexicon centers on verbs that imply controlled engagement with the prompt — confirm, verify, check, ensure, make sure — and on framings that position the candidate as an active participant in the comprehension process rather than as a passive recipient. The deficiency-signaling lexicon centers on verbs that imply uncontrolled failure — hear, catch, understand, follow — and on framings that position the candidate as a failed recipient who needs the prompt to be repeated. The lexical contrast produces a measurable difference in the rater's scoring even when the structural content of the request is identical; the competence-signaling formulation (I want to confirm the specific aspect of the question) and the deficiency-signaling formulation (I did not catch the specific aspect of the question) are propositionally equivalent but are rubric-scored differently because the lexicon triggers different rater inferences about the candidate's discourse competence.
The installation discipline should treat the competence-signaling lexicon as a fixed inventory that the candidate deploys without improvisation during the test, with the lexicon being expanded only after the four-step sequence is fully automated. The candidate who attempts to improvise the lexicon under test conditions frequently defaults to the deficiency-signaling formulations because those formulations are more conversationally natural in Japanese-English code-switching patterns; the fixed inventory is the cognitive scaffolding that prevents the default and maintains the competence-signaling discipline throughout the test session.
The four-week installation drill
The repair-initiation discipline is acquired through a four-week installation drill that progressively builds the four-step sequence into the candidate's speaking automaticity under uncertainty. The drill structure isolates the lexical inventory in week one, integrates the four steps under controlled uncertainty in week two, integrates the move into full-section practice in week three, and pressure-tests the integrated move against high-uncertainty prompts in week four.
Week one — Lexical inventory automation
In week one, the candidate memorizes the competence-signaling lexicon and the recommended signal patterns for each of the four steps, and produces twenty practice responses for each step in isolation. The candidate does not practice the full sequence in week one; the week-one drill isolates the lexical layer so that the lexicon is automatic before the sequence-level practice begins. The pass criterion for week one is the production of the recommended signal patterns within one second of the prompt cue, with the deficiency-signaling lexicon absent from the responses across the full set of practice trials.
Week two — Controlled-uncertainty sequence integration
In week two, the candidate practices the full four-step sequence on ten artificially obscured prompts per day, where the obscuration is engineered to produce a genuine comprehension gap that the candidate must repair. The artificial obscuration can be produced by recording the candidate's practice prompts at reduced volume, by introducing background noise that masks specific prompt segments, or by selecting prompts that combine unfamiliar vocabulary with high-density discourse markers. The week-two drill builds the sequence-level execution under the realistic uncertainty conditions that produce the comprehension gaps during the actual test. The pass criterion for week two is the production of full four-step sequences that pass a self-audit checklist confirming that all four steps are present, that the lexical choices are drawn from the competence-signaling inventory, and that the post-clarification response addresses the clarified prompt rather than the originally guessed prompt.
Week three — Full-section integration
In week three, the candidate practices full TOEIC Link speaking sections under untimed conditions, deploying the repair-initiation move only when a genuine comprehension gap is encountered. The week-three drill is where the candidate develops the discrimination between the prompts that require repair and the prompts that can be responded to directly; the candidate who deploys the repair move on every prompt frequently scores below the candidate who deploys the move only on the prompts where the move is necessary because the over-deployment is scored as a discourse-competence over-claim. The pass criterion for week three is the production of complete sections in which the repair-initiation move is deployed on between fifteen and twenty-five percent of prompts and is absent from the prompts where the candidate has confidently comprehended the question.
Week four — Pressure testing
In week four, the candidate practices complete sections under timed conditions matching the actual TOEIC Link speaking section constraints, and reviews the post-test recordings for repair-initiation execution failures. The error-analysis discipline isolates the residual failure modes — steps that are skipped under time pressure, lexical choices that default to deficiency-signaling formulations, post-clarification responses that revert to the originally guessed prompt — and produces the targeted revision plan that closes the final accuracy gap. The pass criterion for week four is a stable production rate of complete sections that deploy the four-step sequence correctly on between fifteen and twenty-five percent of prompts within the time constraint, with the execution failure rate below fifteen percent.
How the repair-initiation move compounds with other speaking sub-skills
The repair-initiation move is the discourse-competence component that unlocks the band-25-and-above ceiling, but the move's full yield depends on its integration with three other speaking sub-skills. The first is strategic pausing — the four-step sequence requires the candidate to use the response-window opening for the structured repair rather than for silent processing, and the pause discipline determines whether the opening is converted into the rubric-rewarded structure. The second is self-correction-and-repair-strategies competence — the post-clarification response frequently requires real-time adjustments that the broader repair-and-correction repertoire supports, and the candidate who has installed both the repair-initiation move and the broader repair repertoire produces responses that the rater scores as discourse-competent across multiple structural dimensions. The third is paraphrase-and-vocabulary-substitution competence — the comprehension-articulation step requires the candidate to paraphrase the partially comprehended prompt, and the paraphrase competence determines whether the articulation reads as competent partial comprehension or as confused approximation. For the related compound skills, see the speaking paraphrase and vocabulary substitution guide.
The compounding effect is what produces the observed pattern in which candidates who install the repair-initiation move alongside the three supporting sub-skills gain three-to-four band points on the speaking section, while candidates who install the move without the supporting sub-skills gain one band point and frequently regress on the prompts where the over-deployment penalty surfaces. The integrated installation is the operational discipline that produces the full band gain; the isolated installation is the partial discipline that produces the partial gain.
The audit checkpoint
The candidate who has installed the repair-initiation discipline should be able to pass three audit checkpoints that distinguish the installed pattern from the partial implementation that produces no band gain.
First, on a set of ten artificially obscured prompts under timed conditions, the candidate should produce repair-initiation moves on the prompts that contain genuine comprehension gaps and direct responses on the prompts that contain comprehension that the candidate has manufactured. The audit confirms that the discrimination layer is built and that the candidate is not over-deploying the repair move on the prompts where direct response is the rubric-rewarded strategy.
Second, the candidate should produce the four-step sequence on the gap-prompts with the lexical inventory drawn entirely from the competence-signaling lexicon, with no deficiency-signaling formulations across the full set of trials. The audit confirms that the lexical discipline is built and that the candidate is not regressing to the conversationally natural deficiency-signaling formulations under time pressure.
Third, the post-clarification responses should address the clarified prompts rather than the originally guessed prompts, with the response content demonstrating that the clarification has been integrated into the candidate's response construction. The audit confirms that the re-entry step is operationally connected to the candidate's response-generation process and that the repair-initiation move is producing the comprehension correction that justifies the discourse-competence credit.
The candidate who passes the three audits has installed the repair-initiation discipline as a stable component of the speaking repertoire and has converted the prompts with comprehension gaps from a band-22 ceiling enforcement mechanism into the discourse-competence evidence that distinguishes band-25-and-above performance from the band-22 baseline.